In evaluation of wind-induced responses through time-history analysis based on stationary random process, its responsive value must be dealt with numerous numbers of statistics. However, the designing value is often determined from the small number of samples because of limited time and expense in practice. Therefore, it is important and necessary to grasp dispersion of wind response from small number of samples.
In this paper, the dispersion of wind forces and wind-induced responses are investigated from wind tunnel tests. Then, the influence on wind responses due to the difference in the number of samples used as an ensemble average is elucidated.